现代 JavaScript 网络请求库选型指南：Axios 与 Fetch 实现对比

A light-weight module that brings Fetch API to Node.js.

Consider supporting us on our Open Collective:

You might be looking for the v2 docs

• Motivation
• Features
• Difference from client-side fetch
• Installation
• Loading and configuring the module
• Upgrading
• Common Usage
  • Plain text or HTML
  • JSON
  • Simple Post
  • Post with JSON
  • Post with form parameters
  • Handling exceptions
  • Handling client and server errors
  • Handling cookies
• Advanced Usage
  • Streams
  • Accessing Headers and other Metadata
  • Extract Set-Cookie Header
  • Post data using a file
  • Request cancellation with AbortSignal
• API
  • fetch(url[, options])
  • Options
    • Default Headers
    • Custom Agent
    • Custom highWaterMark
    • Insecure HTTP Parser
  • Class: Request
    • new Request(input[, options])
  • Class: Response
    • new Response([body[, options]])
    • response.ok
    • response.redirected
    • response.type
  • Class: Headers
    • new Headers([init])
  • Interface: Body
    • body.body
    • body.bodyUsed
    • body.arrayBuffer()
    • body.blob()
    • body.formData()
    • body.json()
    • body.text()
  • Class: FetchError
  • Class: AbortError
• TypeScript
• Acknowledgement
• Team - Former
• License

Motivation

Instead of implementing XMLHttpRequest in Node.js to run browser-specific Fetch polyfill, why not go from native http to fetch API directly? Hence, node-fetch, minimal code for a window.fetch compatible API on Node.js runtime.

See Jason Miller's isomorphic-unfetch or Leonardo Quixada's cross-fetch for isomorphic usage (exports node-fetch for server-side, whatwg-fetch for client-side).

Features

• Stay consistent with window.fetch API.
• Make conscious trade-off when following WHATWG fetch spec and stream spec implementation details, document known differences.
• Use native promise and async functions.
• Use native Node streams for body, on both request and response.
• Decode content encoding (gzip/deflate/brotli) properly, and convert string output (such as res.text() and res.json()) to UTF-8 automatically.
• Useful extensions such as redirect limit, response size limit, explicit errors for troubleshooting.

Difference from client-side fetch

• See known differences:
  • As of v3.x
  • As of v2.x
• If you happen to use a missing feature that window.fetch offers, feel free to open an issue.
• Pull requests are welcomed too!

Installation

Current stable release (3.x) requires at least Node.js 12.20.0.

npm install node-fetch

Loading and configuring the module

ES Modules (ESM)

import fetch from 'node-fetch';

CommonJS

node-fetch from v3 is an ESM-only module - you are not able to import it with require().

If you cannot switch to ESM, please use v2 which remains compatible with CommonJS. Critical bug fixes will continue to be published for v2.

npm install node-fetch@2

Alternatively, you can use the async import() function from CommonJS to load node-fetch asynchronously:

// mod.cjs
const fetch = (...args) => import('node-fetch').then(({default: fetch}) => fetch(...args));

Providing global access

To use fetch() without importing it, you can patch the global object in node:

// fetch-polyfill.js
import fetch, {
  Blob,
  blobFrom,
  blobFromSync,
  File,
  fileFrom,
  fileFromSync,
  FormData,
  Headers,
  Request,
  Response,
} from 'node-fetch'

if (!globalThis.fetch) {
  globalThis.fetch = fetch
  globalThis.Headers = Headers
  globalThis.Request = Request
  globalThis.Response = Response
}

// index.js
import './fetch-polyfill'

// ...

Upgrading

Using an old version of node-fetch? Check out the following files:

• 2.x to 3.x upgrade guide
• 1.x to 2.x upgrade guide
• Changelog

Common Usage

NOTE: The documentation below is up-to-date with 3.x releases, if you are using an older version, please check how to upgrade.

Plain text or HTML

import fetch from 'node-fetch';

const response = await fetch('https://github.com/');
const body = await response.text();

console.log(body);

JSON

import fetch from 'node-fetch';

const response = await fetch('https://api.github.com/users/github');
const data = await response.json();

console.log(data);

Simple Post

import fetch from 'node-fetch';

const response = await fetch('https://httpbin.org/post', {method: 'POST', body: 'a=1'});
const data = await response.json();

console.log(data);

Post with JSON

import fetch from 'node-fetch';

const body = {a: 1};

const response = await fetch('https://httpbin.org/post', {
	method: 'post',
	body: JSON.stringify(body),
	headers: {'Content-Type': 'application/json'}
});
const data = await response.json();

console.log(data);

Post with form parameters

URLSearchParams is available on the global object in Node.js as of v10.0.0. See official documentation for more usage methods.

NOTE: The Content-Type header is only set automatically to x-www-form-urlencoded when an instance of URLSearchParams is given as such:

import fetch from 'node-fetch';

const params = new URLSearchParams();
params.append('a', 1);

const response = await fetch('https://httpbin.org/post', {method: 'POST', body: params});
const data = await response.json();

console.log(data);

Handling exceptions

NOTE: 3xx-5xx responses are NOT exceptions, and should be handled in then(), see the next section.

Wrapping the fetch function into a try/catch block will catch all exceptions, such as errors originating from node core libraries, like network errors, and operational errors which are instances of FetchError. See the error handling document for more details.

import fetch from 'node-fetch';

try {
	await fetch('https://domain.invalid/');
} catch (error) {
	console.log(error);
}

Handling client and server errors

It is common to create a helper function to check that the response contains no client (4xx) or server (5xx) error responses:

import fetch from 'node-fetch';

class HTTPResponseError extends Error {
	constructor(response) {
		super(`HTTP Error Response: ${response.status} ${response.statusText}`);
		this.response = response;
	}
}

const checkStatus = response => {
	if (response.ok) {
		// response.status >= 200 && response.status < 300
		return response;
	} else {
		throw new HTTPResponseError(response);
	}
}

const response = await fetch('https://httpbin.org/status/400');

try {
	checkStatus(response);
} catch (error) {
	console.error(error);

	const errorBody = await error.response.text();
	console.error(`Error body: ${errorBody}`);
}

Handling cookies

Cookies are not stored by default. However, cookies can be extracted and passed by manipulating request and response headers. See Extract Set-Cookie Header for details.

Advanced Usage

Streams

The "Node.js way" is to use streams when possible. You can pipe res.body to another stream. This example uses stream.pipeline to attach stream error handlers and wait for the download to complete.

import {createWriteStream} from 'node:fs';
import {pipeline} from 'node:stream';
import {promisify} from 'node:util'
import fetch from 'node-fetch';

const streamPipeline = promisify(pipeline);

const response = await fetch('https://github.githubassets.com/images/modules/logos_page/Octocat.png');

if (!response.ok) throw new Error(`unexpected response ${response.statusText}`);

await streamPipeline(response.body, createWriteStream('./octocat.png'));

In Node.js 14 you can also use async iterators to read body; however, be careful to catch errors -- the longer a response runs, the more likely it is to encounter an error.

import fetch from 'node-fetch';

const response = await fetch('https://httpbin.org/stream/3');

try {
	for await (const chunk of response.body) {
		console.dir(JSON.parse(chunk.toString()));
	}
} catch (err) {
	console.error(err.stack);
}

In Node.js 12 you can also use async iterators to read body; however, async iterators with streams did not mature until Node.js 14, so you need to do some extra work to ensure you handle errors directly from the stream and wait on it response to fully close.

import fetch from 'node-fetch';

const read = async body => {
	let error;
	body.on('error', err => {
		error = err;
	});

	for await (const chunk of body) {
		console.dir(JSON.parse(chunk.toString()));
	}

	return new Promise((resolve, reject) => {
		body.on('close', () => {
			error ? reject(error) : resolve();
		});
	});
};

try {
	const response = await fetch('https://httpbin.org/stream/3');
	await read(response.body);
} catch (err) {
	console.error(err.stack);
}

Accessing Headers and other Metadata

import fetch from 'node-fetch';

const response = await fetch('https://github.com/');

console.log(response.ok);
console.log(response.status);
console.log(response.statusText);
console.log(response.headers.raw());
console.log(response.headers.get('content-type'));

Extract Set-Cookie Header

Unlike browsers, you can access raw Set-Cookie headers manually using Headers.raw(). This is a node-fetch only API.

import fetch from 'node-fetch';

const response = await fetch('https://example.com');

// Returns an array of values, instead of a string of comma-separated values
console.log(response.headers.raw()['set-cookie']);

Post data using a file

import fetch, {
  Blob,
  blobFrom,
  blobFromSync,
  File,
  fileFrom,
  fileFromSync,
} from 'node-fetch'

const mimetype = 'text/plain'
const blob = fileFromSync('./input.txt', mimetype)
const url = 'https://httpbin.org/post'

const response = await fetch(url, { method: 'POST', body: blob })
const data = await response.json()

console.log(data)

node-fetch comes with a spec-compliant FormData implementations for posting multipart/form-data payloads

import fetch, { FormData, File, fileFrom } from 'node-fetch'

const httpbin = 'https://httpbin.org/post'
const formData = new FormData()
const binary = new Uint8Array([ 97, 98, 99 ])
const abc = new File([binary], 'abc.txt', { type: 'text/plain' })

formData.set('greeting', 'Hello, world!')
formData.set('file-upload', abc, 'new name.txt')

const response = await fetch(httpbin, { method: 'POST', body: formData })
const data = await response.json()

console.log(data)

If you for some reason need to post a stream coming from any arbitrary place, then you can append a Blob or a File look-a-like item.

The minimum requirement is that it has:

1. A Symbol.toStringTag getter or property that is either Blob or File
2. A known size.
3. And either a stream() method or a arrayBuffer() method that returns a ArrayBuffer.

The stream() must return any async iterable object as long as it yields Uint8Array (or Buffer) so Node.Readable streams and whatwg streams works just fine.

formData.append('upload', {
	[Symbol.toStringTag]: 'Blob',
	size: 3,
  *stream() {
    yield new Uint8Array([97, 98, 99])
	},
	arrayBuffer() {
		return new Uint8Array([97, 98, 99]).buffer
	}
}, 'abc.txt')

Request cancellation with AbortSignal

You may cancel requests with AbortController. A suggested implementation is abort-controller.

An example of timing out a request after 150ms could be achieved as the following:

import fetch, { AbortError } from 'node-fetch';

// AbortController was added in node v14.17.0 globally
const AbortController = globalThis.AbortController || await import('abort-controller')

const controller = new AbortController();
const timeout = setTimeout(() => {
	controller.abort();
}, 150);

try {
	const response = await fetch('https://example.com', {signal: controller.signal});
	const data = await response.json();
} catch (error) {
	if (error instanceof AbortError) {
		console.log('request was aborted');
	}
} finally {
	clearTimeout(timeout);
}

See test cases for more examples.

API

fetch(url[, options])

• url A string representing the URL for fetching
• options Options for the HTTP(S) request
• Returns: Promise<Response>

Perform an HTTP(S) fetch.

url should be an absolute URL, such as https://example.com/. A path-relative URL (/file/under/root) or protocol-relative URL (//can-be-http-or-https.com/) will result in a rejected Promise.

Options

The default values are shown after each option key.

{
	// These properties are part of the Fetch Standard
	method: 'GET',
	headers: {},            // Request headers. format is the identical to that accepted by the Headers constructor (see below)
	body: null,             // Request body. can be null, or a Node.js Readable stream
	redirect: 'follow',     // Set to `manual` to extract redirect headers, `error` to reject redirect
	signal: null,           // Pass an instance of AbortSignal to optionally abort requests

	// The following properties are node-fetch extensions
	follow: 20,             // maximum redirect count. 0 to not follow redirect
	compress: true,         // support gzip/deflate content encoding. false to disable
	size: 0,                // maximum response body size in bytes. 0 to disable
	agent: null,            // http(s).Agent instance or function that returns an instance (see below)
	highWaterMark: 16384,   // the maximum number of bytes to store in the internal buffer before ceasing to read from the underlying resource.
	insecureHTTPParser: false	// Use an insecure HTTP parser that accepts invalid HTTP headers when `true`.
}

Default Headers

If no values are set, the following request headers will be sent automatically:

Note: when body is a Stream, Content-Length is not set automatically.

Custom Agent

The agent option allows you to specify networking related options which are out of the scope of Fetch, including and not limited to the following:

• Support self-signed certificate
• Use only IPv4 or IPv6
• Custom DNS Lookup

See http.Agent for more information.

If no agent is specified, the default agent provided by Node.js is used. Note that this changed in Node.js 19 to have keepalive true by default. If you wish to enable keepalive in an earlier version of Node.js, you can override the agent as per the following code sample.

In addition, the agent option accepts a function that returns http(s).Agent instance given current URL, this is useful during a redirection chain across HTTP and HTTPS protocol.

import http from 'node:http';
import https from 'node:https';

const httpAgent = new http.Agent({
	keepAlive: true
});
const httpsAgent = new https.Agent({
	keepAlive: true
});

const options = {
	agent: function(_parsedURL) {
		if (_parsedURL.protocol == 'http:') {
			return httpAgent;
		} else {
			return httpsAgent;
		}
	}
};

Custom highWaterMark

Stream on Node.js have a smaller internal buffer size (16kB, aka highWaterMark) from client-side browsers (>1MB, not consistent across browsers). Because of that, when you are writing an isomorphic app and using res.clone(), it will hang with large response in Node.

The recommended way to fix this problem is to resolve cloned response in parallel:

import fetch from 'node-fetch';

const response = await fetch('https://example.com');
const r1 = response.clone();

const results = await Promise.all([response.json(), r1.text()]);

console.log(results[0]);
console.log(results[1]);

If for some reason you don't like the solution above, since 3.x you are able to modify the highWaterMark option:

import fetch from 'node-fetch';

const response = await fetch('https://example.com', {
	// About 1MB
	highWaterMark: 1024 * 1024
});

const result = await res.clone().arrayBuffer();
console.dir(result);

Insecure HTTP Parser

Passed through to the insecureHTTPParser option on http(s).request. See http.request for more information.

Manual Redirect

The redirect: 'manual' option for node-fetch is different from the browser & specification, which results in an opaque-redirect filtered response. node-fetch gives you the typical basic filtered response instead.

import fetch from 'node-fetch';

const response = await fetch('https://httpbin.org/status/301', { redirect: 'manual' });

if (response.status === 301 || response.status === 302) {
	const locationURL = new URL(response.headers.get('location'), response.url);
	const response2 = await fetch(locationURL, { redirect: 'manual' });
	console.dir(response2);
}

Class: Request

An HTTP(S) request containing information about URL, method, headers, and the body. This class implements the Body interface.

Due to the nature of Node.js, the following properties are not implemented at this moment:

• type
• destination
• mode
• credentials
• cache
• integrity
• keepalive

The following node-fetch extension properties are provided:

• follow
• compress
• counter
• agent
• highWaterMark

See options for exact meaning of these extensions.

new Request(input[, options])

(spec-compliant)

• input A string representing a URL, or another Request (which will be cloned)
• options Options for the HTTP(S) request

Constructs a new Request object. The constructor is identical to that in the browser.

In most cases, directly fetch(url, options) is simpler than creating a Request object.

Class: Response

An HTTP(S) response. This class implements the Body interface.

The following properties are not implemented in node-fetch at this moment:

• trailer

new Response([body[, options]])

(spec-compliant)

• body A String or Readable stream
• options A ResponseInit options dictionary

Constructs a new Response object. The constructor is identical to that in the browser.

Because Node.js does not implement service workers (for which this class was designed), one rarely has to construct a Response directly.

response.ok

(spec-compliant)

Convenience property representing if the request ended normally. Will evaluate to true if the response status was greater than or equal to 200 but smaller than 300.

response.redirected

(spec-compliant)

Convenience property representing if the request has been redirected at least once. Will evaluate to true if the internal redirect counter is greater than 0.

response.type

(deviation from spec)

Convenience property representing the response's type. node-fetch only supports 'default' and 'error' and does not make use of filtered responses.

Class: Headers

This class allows manipulating and iterating over a set of HTTP headers. All methods specified in the Fetch Standard are implemented.

new Headers([init])

(spec-compliant)

• init Optional argument to pre-fill the Headers object

Construct a new Headers object. init can be either null, a Headers object, an key-value map object or any iterable object.

// Example adapted from https://fetch.spec.whatwg.org/#example-headers-class
import {Headers} from 'node-fetch';

const meta = {
	'Content-Type': 'text/xml'
};
const headers = new Headers(meta);

// The above is equivalent to
const meta = [['Content-Type', 'text/xml']];
const headers = new Headers(meta);

// You can in fact use any iterable objects, like a Map or even another Headers
const meta = new Map();
meta.set('Content-Type', 'text/xml');
const headers = new Headers(meta);
const copyOfHeaders = new Headers(headers);

Interface: Body

Body is an abstract interface with methods that are applicable to both Request and Response classes.

body.body

(deviation from spec)

• Node.js Readable stream

Data are encapsulated in the Body object. Note that while the Fetch Standard requires the property to always be a WHATWG ReadableStream, in node-fetch it is a Node.js Readable stream.

body.bodyUsed

(spec-compliant)

• Boolean

A boolean property for if this body has been consumed. Per the specs, a consumed body cannot be used again.

body.arrayBuffer()

body.formData()

body.blob()

body.json()

body.text()

fetch comes with methods to parse multipart/form-data payloads as well as x-www-form-urlencoded bodies using .formData() this comes from the idea that Service Worker can intercept such messages before it's sent to the server to alter them. This is useful for anybody building a server so you can use it to parse & consume payloads.

import http from 'node:http'
import { Response } from 'node-fetch'

http.createServer(async function (req, res) {
  const formData = await new Response(req, {
    headers: req.headers // Pass along the boundary value
  }).formData()
  const allFields = [...formData]

  const file = formData.get('uploaded-files')
  const arrayBuffer = await file.arrayBuffer()
  const text = await file.text()
  const whatwgReadableStream = file.stream()

  // other was to consume the request could be to do:
  const json = await new Response(req).json()
  const text = await new Response(req).text()
  const arrayBuffer = await new Response(req).arrayBuffer()
  const blob = await new Response(req, {
    headers: req.headers // So that `type` inherits `Content-Type`
  }.blob()
})

Class: FetchError

(node-fetch extension)

An operational error in the fetching process. See ERROR-HANDLING.md for more info.

Class: AbortError

(node-fetch extension)

An Error thrown when the request is aborted in response to an AbortSignal's abort event. It has a name property of AbortError. See ERROR-HANDLING.MD for more info.

TypeScript

Since 3.x types are bundled with node-fetch, so you don't need to install any additional packages.

For older versions please use the type definitions from DefinitelyTyped:

npm install --save-dev @types/node-fetch@2.x

Acknowledgement

Thanks to github/fetch for providing a solid implementation reference.

Team

David Frank	Jimmy Wärting	Antoni Kepinski	Richie Bendall	Gregor Martynus

Former

• Timothy Gu
• Jared Kantrowitz

License

MIT